Ecological-Modeling on Yuval Bloch — Complexity & Ecology

Three lessons on Bayesian optimization, and one on humility

Fri, 17 Jul 2026 00:00:00 +0000

Two weeks ago, I told my advisor that my thesis was ready. We decided to delay the submission a little — partly so he could find a referee, and partly so I could keep my scholarship while I converted the thesis into a paper and made sure my code was reproducible.

Three days ago, I finished building a simple pipeline that would let future researchers reproduce all of my results without needing me around. I ran a quick test to confirm that I could reproduce everything without diving back into the code myself — and when I ran a full reproduction through the pipeline, I got completely different results.

End of Greatness: Lacunarity and the Index of Transitional Homogeneity

Fri, 20 Feb 2026 00:00:00 +0000

One of the best things about being an ecologist in this era is the constant opportunity to bridge different disciplines. Recently, my research led me to use a measure called transitional homogeneity—a concept originally developed in cosmology to estimate how matter is spread throughout the universe. My journey to this point was a classic scientific process of trial and error.

The Quest for a Multi-Scale Metric

After building a cellular automata model to generate future land-use maps, I needed a way to quantify the structural differences between various scenarios. Traditional indices, such as edge density (the total length of edges between different land uses), only provide a small part of the picture. They focus on local interactions—representing the contact between two land-use types—without considering the broader landscape context

The Devil Is in the Details: Modeling Small-Scale Land-Use Change in Madagascar

Sun, 07 Dec 2025 00:00:00 +0000

🖼️ The Poster

If you are viewing this on a mobile device, you can view the full PDF below or Download the PDF here.

📌 Introduction & Research Gap

Land use represents the fundamental interface between humanity and nature. From urban centers to remote tropical regions, environmental modification critically affects human health, disease exposure, and vulnerability to natural disasters.

Land-use change is a primary driver of global crises:

Explaining the Lacunarity-Based Index for Spatial Heterogeneity

Wed, 22 Oct 2025 00:00:00 +0000

In my research, I employ the measurement of lacunarity based on the methodology described in Scott, R., et al. “A Lacunarity-based Index for Spatial Heterogeneity.” Earth and Space Science (Hoboken, N.J.), vol. 9, no. 8, American Geophysical Union (AGU), Aug. 2022, https://doi.org/10.1029/2021ea002180.

Here, I will provide a less formal description of the mathematics behind this method and its implementation in Julia.

Note: Since writing this post, subsequent analysis revealed a strong correlation between Scott’s (R) measure and edge density. To address this issue, I switched to the Index of Transitional Homogeneity, which provides a more appropriate measure of spatial structure.

Sensitivity to Land Configuration

Mon, 07 Jul 2025 10:00:00 +0300

Introduction

Natural populations are influenced not only by the amount of available habitat but also by its spatial configuration. Factors such as edge effects, habitat fragmentation, and patch isolation are critical in shaping population dynamics. These spatial processes are fundamental in ecology, and it’s essential that my model captures them accurately.

While the total amount of habitat is relatively easy to quantify, capturing spatial configuration is far more complex. Traditional models often rely on a binary landscape, classifying areas as either suitable or unsuitable for the organism. Metrics like edge length, connectivity, and mean patch size are often used in these models. However, in my system, the landscape consists of multiple land cover types, each offering varying degrees of suitability for ticks. This complexity demands a more nuanced and flexible approach.

Finding Viable Parameter Spaces with Sensitivity Analysis

Mon, 09 Jun 2025 10:00:00 +0300

Introduction

Ecological models often depend on parameters that are difficult to measure precisely. Sensitivity analysis helps us ask: how much do these uncertainties matter? It allows us to explore:

Which parameters most strongly influence model outcomes
What combinations lead to persistence or extinction
How robust the system is to changes in key assumptions

In this project, I go beyond simply measuring sensitivity. I use these tools to map the boundary between viable and non-viable ecological conditions—the tipping points where tick populations either collapse or persist. By comparing these boundaries to patterns observed in real-world data, I can ask a deeper question: which combinations of parameters are not just viable, but actually relevant to the system I’m trying to model?

Simulating Tick Life Cycles Across Landscapes

Thu, 05 Jun 2025 18:31:05 +0300

Model

This model simulates tick population dynamics using a 3-layer lattice, where each layer corresponds to a life stage: larva, nymph, and adult. Ticks move both vertically between life stages and horizontally between geographic locations by biting a host.

Lattice and Tick Population

Each land cell is denoted by $L_{i,j}$, with a tick-carrying capacity $K_{i,j}$.
The population at life stage $s$ in cell $L_{i,j}$ is represented by $P_{i,j,s}$.

Figure 1. Each life stage is represented as a 2D lattice layer. Bites cause vertical (stage) and horizontal (spatial) movement.

Tick Movement and Life Cycle

To simplify, the model is first described in discrete steps: